
This repository presents a comprehensive computer vision methodology for automated crack detection in concrete cultural heritage structures, applied to Ghana's Independence Arch (Accra). The study addresses critical conservation challenges for marine-exposed monuments through UAV-based image acquisition and deep learning classification. Key contributions include: site-specific environmental analysis (extreme marine exposure per GB 5009-2012, ASCE-7 wind loading); Convolutional Neural Network architecture optimized for concrete crack detection (94% validation accuracy); complete image processing pipeline: augmentation, normalization, whitening, and binary classification; practical damage control framework integrating detection with repair strategies (epoxy injection, self-healing concrete); economic impact assessment linking structural conservation to tourism revenue. The methodology demonstrates scalable, low-cost structural health monitoring for developing nation heritage infrastructure.